Dementia occurs in up to 80% of people with idiopathic Parkinson disease (PD) and substantially increases morbidity, mortality and cost of care for these individuals. In fact, it may be the most important factor leading to residential care placement. Yet, treatment of dementia in PD remains nearly non-existent, making this one of the great unmet needs for the nearly one million people affected by PD in North America. A major roadblock in developing and testing new treatments for dementia in PD is identification of the cause of dementia in a given individual - critically important to test therapies in relatively homogenous groups. This is a particular challenge in PD since dementia may be caused by either underlying cortical a-synuclein (a-syn) pathology or co-existing Alzheimer's (AD) pathology that includes abnormal deposition of A and tau proteins. Our recent studies suggest that this may be an oversimplification. A may be an independent factor, not necessarily related to tauopathy in PD, producing 3 groups of PD dementia: 1) primarily cortical synucleinopathy, 2) primarily cortical synucleinopathy with abnormal A and 3) much less commonly cortical synucleinopathy with abnormal A and tauopathy. We will test the hypotheses that behavior, PET, MRI, genetic and CSF biomarkers will permit identification of these causes and help predict onset of dementia in PD. We propose to test these hypotheses with a longitudinal follow up of people with PD with and without dementia and matched healthy controls. The multidisciplinary team will do evaluations including standard clinical evaluation, clinical dementia ratings, neuropsychological testing, volumetric MRI based measures of regional atrophy, resting state functional connectivity MR of brain networks, PET PIB in vivo measures of amyloid, CSF measures of relevant proteins (a-syn, A42 and tau), postmortem neuropathology and quantified analysis of a-syn, A and tau in selected brain areas. The postmortem analyses of brains will permit validation of the biomarkers. This also will provide the opportunity to make connections between imaging findings, behavior and regional protein expression in the brain. This study will validate biomarkers of specific pathologies underlying dementia in PD and that predict dementia onset. Our approach, borrowed heavily from studies in AD, will provide the framework for testing to-be-developed measures such as new CSF proteins, imaging methods or genetic signatures to determine whether they are more reliable or earlier biomarkers of dementia in PD. The ultimate goal is to use these biomarkers to test the efficacy of new therapies that will prevent dementia in PD.